Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Congestion Control in Charging Stations Allocation with Q-Learning
Navigation systems can help in allocating public charging stations to electric vehicles (EVs) with the aim of minimizing EVs’ charging time by integrating sufficient data. However, the existing systems only consider their travel time and transform the allocation as a routing problem. In this paper, we involve the queuing time in stations as one part of EVs’ charging time, and another part is the travel time on roads. Roads and stations are easily congested resources, and we constructed a joint-resource congestion game to describe the interaction between vehicles and resources. With a finite number of vehicles and resources, there exists a Nash equilibrium. To realize a self-adaptive allocation work, we applied the Q-learning algorithm on systems, defining sets of states and actions in our constructed environment. After being allocated one by one, vehicles concurrently requesting to be charged will be processed properly. We collected urban road network data from Chongqing city and conducted experiments. The results illustrate the proposed method can be used to solve the problem, and its convergence performance was better than the genetic algorithm. The road capacity and the number of EVs affected the initial of Q-value, and not the convergence trends.
Congestion Control in Charging Stations Allocation with Q-Learning
Navigation systems can help in allocating public charging stations to electric vehicles (EVs) with the aim of minimizing EVs’ charging time by integrating sufficient data. However, the existing systems only consider their travel time and transform the allocation as a routing problem. In this paper, we involve the queuing time in stations as one part of EVs’ charging time, and another part is the travel time on roads. Roads and stations are easily congested resources, and we constructed a joint-resource congestion game to describe the interaction between vehicles and resources. With a finite number of vehicles and resources, there exists a Nash equilibrium. To realize a self-adaptive allocation work, we applied the Q-learning algorithm on systems, defining sets of states and actions in our constructed environment. After being allocated one by one, vehicles concurrently requesting to be charged will be processed properly. We collected urban road network data from Chongqing city and conducted experiments. The results illustrate the proposed method can be used to solve the problem, and its convergence performance was better than the genetic algorithm. The road capacity and the number of EVs affected the initial of Q-value, and not the convergence trends.
Congestion Control in Charging Stations Allocation with Q-Learning
Li Zhang (Autor:in) / Ke Gong (Autor:in) / Maozeng Xu (Autor:in)
2019
Aufsatz (Zeitschrift)
Elektronische Ressource
Unbekannt
Metadata by DOAJ is licensed under CC BY-SA 1.0
LONDON CONGESTION CHARGING SCHEME
| British Library Conference Proceedings | 2004
| Online Contents | 2008
Congestion Charging for London
| British Library Online Contents | 1994
| Springer Verlag | 2022
Congestion Charging - das Beispiel London
| IuD Bahn | 2004